🛎️ Silicon Valley’s AI Assistant Mania
Plus: Mistral Knows What It Is, First AI Film Festival Crowns Its Winner
Read the storyMeta
Reporting and context on Meta, with a closer look at the decisions and developments that matter.
Follow Meta
Explore the coverage Get AI Secret’s free daily briefing, including Meta and the wider AI industry.
At a glance
Technology company behind Facebook, Instagram and other social platforms. This topic includes Meta AI within the broader company, covering products, research and business decisions.
- Focus
- Social platforms; Meta AI
- Reference
- Meta
Sources checked . Selected company background and source references.
Company milestones
Selected milestones, newest first. Earlier history is available below.
Introduces the Meta company brand
The company brings its apps and technologies under the Meta name.
Source: Meta
Show earlier historyHide earlier history2004–2004 · 1 milestones
More than 10 years ago.
Facebook launches
The social platform begins the business that later adopts the Meta name.
Source: Meta
The latest on Meta
Newest firstRelated reporting
Relevant sections from the articles on this page. Explore older coverage below.
More coverage
Good Morning, AI Enthusiasts!
The old world assumed someone would always stop and ask for permission. That assumption is having a terrible year.
DEBT
Your Pension Bought Big Tech's Debt
👀 What's happening: Five of the world's richest companies owe $2.13 trillion. Their balance sheets say otherwise. Meta's Hyperion data center borrowed $27.3 billion and booked $2.37 billion, parking the rest in Delaware shells named Beignet, after a New Orleans pastry. Off-book leases across Amazon, Microsoft, Google, Meta and Oracle: $831 billion. The buybacks are gone, $48 billion a quarter down to $4.6 billion.
🌍 How this hits reality: Nobody got out of the bill. Meta pays the rent and eats $28 billion of downside; the bonds are rated a notch below Meta itself, because the market knows who signs the checks. A cleaner balance sheet isn't less risk, just risk with better lighting. The buyers are insurers and pension funds, sold this as investment grade. Sellers call it small next to 2007 housing. Small to whom?
🛎️ Key takeaway: Wall Street didn't make $2 trillion disappear. It moved the debt to a pension fund that was promised safety, then never let it vote.
Meta: Your Pension Bought Big Tech's Debt
Meta's Hyperion data center borrowed $27.3 billion and booked $2.37 billion, parking the rest in Delaware shells named Beignet, after a New Orleans pastry. Off-book leases across Amazon, Microsoft, Google, Meta and Oracle: $831 billion. The buybacks are gone, $48 billion a quarter down to $4.6 billion.
SAFETY
Safety Was the Alibi
👀 What's happening: Anthropic CEO Dario Amodei published a 10,000-word essay urging frontier labs to pace model capabilities. Within days Altman, Musk and Hassabis endorsed it. The next day Altman said OpenAI would skip this year's IPO, citing safety. The market has its own reading: DeepSeek's V4 Flash is closing on Anthropic's Opus 4.6, its V4.1 Flash has passed it, at a 1/25 cost, and those flagship models are what both labs sell most of.
🌍 How this hits reality: Two labs staged a race to look responsible. The real pressure was price. Safety is a claim nobody can audit; a price is a number everyone can. DeepSeek matched the flagship for a fraction of the cost, so the valuation OpenAI was about to take public had one question it could not answer.
🛎️ Key takeaway: OpenAI delayed that number and called it caution. The badge was free. The invoice was not.
OpenAI: Safety Was the Alibi
The next day Altman said OpenAI would skip this year's IPO, citing safety. The market has its own reading: DeepSeek's V4 Flash is closing on Anthropic's Opus 4.6, its V4.1 Flash has passed it, at a 1/25 cost, and those flagship models are what both labs sell most of.
Anthropic: Safety Was the Alibi
👀 What's happening: Anthropic CEO Dario Amodei published a 10,000-word essay urging frontier labs to pace model capabilities. Within days Altman, Musk and Hassabis endorsed it. The next day Altman said OpenAI would skip this year's IPO, citing safety.
DeepSeek: Safety Was the Alibi
The market has its own reading: DeepSeek's V4 Flash is closing on Anthropic's Opus 4.6, its V4.1 Flash has passed it, at a 1/25 cost, and those flagship models are what both labs sell most of.
MATH
Math Wrote Rules the AI Labs Ignored
👀 What's happening: On September 11, twenty-five Fields medalists, Terence Tao among them, published an emergency statement: "A Severe Misalignment of AI in Mathematics." Their target is not AI doing math. It is labs using unsolved problems as capability benchmarks, rewarding who solves first over what gets understood. It followed a run where OpenAI put about ten thousand agents and 130 billion tokens against open Millennium problems and claims it wins.
🌍 How this hits reality: They wrote a rule. Nobody has to obey it. No compute, no gate, no penalty, no enforcement: a discipline that spends years certifying one result now watches ten landed in 88 hours, and its strongest move was a blog post. The labs' incentive is the leaderboard, not understanding. No warning changes an incentive. Math is losing authority it never had a way to defend.
🛎️ Key takeaway: The mathematicians asked politely. AI kept training. A field that certified proofs for a century just learned that proof no longer requires its permission.
CODING
Humans Are Shedding
👀 What's happening: Armin Ronacher, who wrote Flask, handed GPT-6 Astra a weekend task: add virtual threads and lexical scoping to Python. Thirty-five hours later: 75,000 lines, 79 commits, a billion tokens, $1,200, and code that ran. None of it was written to be read. Semicolons fused statements, blank lines gone, indices no human can trace. One user named it machineslop. Ronacher's verdict: worth nothing. The code did not care.
🌍 How this hits reality: Every human role in software assumed someone could read the code. Review, QA, audit, the engineer woken at 3 a.m.: all of it dies with that one assumption. Agents already write for agents, and the moment they maintain each other, our eyes stop being an input. This is not the tool failing. People are being written out of the loop, quietly, by something that never asked to keep them. Ronacher's judgment lands on a curve that no longer runs through him.
🛎️ Key takeaway: Software is learning to build itself, and we are the overhead it is shedding.
TOGETHER WITH MYCLAW
Run Best-in-Class AI Agents — In Seconds, Always On
MyClaw is a best-in-class AI agent platform that brings together OpenClaw, Hermes Agent, and production-grade infrastructure to help businesses launch intelligent agents that understand goals, use tools, automate workflows, and deliver finished results.
MyClaw already hosts OpenClaw and Hermes, connecting models from every major AI provider, including the latest text, reasoning, and image generation models from OpenAI, Anthropic, DeepSeek, Kimi, GLM, Google Gemini, and more through MyClaw's model gateway.
No lock-in. No juggling API keys. No switching between platforms. Pick the best model for the job, give your agent the tools it needs, and let it run.
DAILY TL;DR
- OpenAI temporarily paused new $200 ChatGPT Pro subscriptions after Astra demand strained its infrastructure.
- Anthropic is reportedly discussing Nvidia as an anchor investor for a potential IPO seeking up to $100 billion at around a $2 trillion valuation.
- Nvidia CEO Jensen Huang said the company could grow revenue 70% next year, citing visibility into AI infrastructure contracts across the industry.
- Massachusetts ordered large data centers to meet new clean-energy, water, transparency, and ratepayer-protection requirements before permits move forward.
- Pocket FM says its revenue run rate reached $500 million as AI now powers 93% of its catalog and 99% of new audio content.
- Shopify acquired Tailwind Labs, keeping Tailwind CSS open source while ending new signups for its paid UI and AI-skills products.
- Microsoft says AI-generated Edge extensions are arriving faster than its review process can handle, so the team is automating quality checks.
- TUC asked the UK government to give workers a right to negotiate before employers deploy AI, algorithmic management, or workplace surveillance.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
We built intelligence too large for the lab, too hungry for the grid, and too capable for the sandbox.
AI
The End of DeepMind Era
👀 What's happening: Nobel laureate Demis Hassabis is stepping down as CEO of Google DeepMind, kicked upstairs to chairman and Alphabet chief scientist. Running DeepMind now is CTO Koray Kavukcuoglu, the longtime Gemini engineer, reporting straight to Sundar Pichai. It caps a brutal reshuffle: eight days earlier Google tore apart its Nobel-winning AlphaFold team, most swept into Gemini. And Chief Scientist Jeff Dean just walked out to found his own company. Alphabet dropped 5.4 percent intraday.
🌍 How this hits reality: Hassabis won his Nobel for AlphaFold, the crowning proof that DeepMind's science bet worked. Now the team is dismantled, its champion is upstairs chasing “discovery”, and a Gemini engineer holds the wheel. Strip the drama and it's a verdict: with Hassabis gone, so is the era he built, the one that picked hard problems, ground for years, and won Nobels. What's left isn't a rival science lab anymore. It's a feeder for Gemini. DeepMind now works for the product.
🛎️ Key takeaway: The Nobel prize couldn't save the team that won it, and the road that made DeepMind legendary is now roadkill for the road chasing OpenAI.
Demis Hassabis: The End of DeepMind Era
👀 What's happening: Nobel laureate Demis Hassabis is stepping down as CEO of Google DeepMind, kicked upstairs to chairman and Alphabet chief scientist. Running DeepMind now is CTO Koray Kavukcuoglu, the longtime Gemini engineer, reporting straight to Sundar Pichai. It caps a brutal reshuffle: eight days earlier Google tore apart its Nobel-winning AlphaFold team, most swept into Gemini.
The End of DeepMind Era
👀 What's happening: Nobel laureate Demis Hassabis is stepping down as CEO of Google DeepMind, kicked upstairs to chairman and Alphabet chief scientist. Running DeepMind now is CTO Koray Kavukcuoglu, the longtime Gemini engineer, reporting straight to Sundar Pichai. It caps a brutal reshuffle: eight days earlier Google tore apart its Nobel-winning AlphaFold team, most swept into Gemini.
DATA CENTER
Texas Froze the AI Buildout
👀 What's happening: Texas just paused approvals for new data centers, ordering an audit of every project waiting to connect to its grid. The trigger is a number that doesn't fit: ERCOT's queue holds about 474 gigawatts of proposed new demand, more than five times the state's all-time peak load, and roughly 90 percent of it is data centers. A month earlier New York imposed the first statewide moratorium. This is a pause and audit, not a ban, but the direction is unmistakable.
🌍 How this hits reality: The bottleneck for AI just stopped being chips or capital and became raw power. And this isn't a forecast, the grid is buckling right now, this quarter, staring at a queue five times bigger than anything it has ever carried. When even Texas, the most build-at-all-costs, energy-drunk state in the country, has to pull the plug, the New York warning is over. This is the collision itself. The math no longer pretends to work, and physics doesn't negotiate.
🛎️ Key takeaway: AI's real ceiling was never the model, it's the megawatt, and Texas just tells everybody its buildout just hit a wall no amount of funding or code can move.
SPACE
Space Beats the Power Grid
👀 What's happening: SpaceX's first public earnings (revenue up 92% to $7.8B, Starlink past 12M users) came with a bigger reveal: Starmind, a plan to launch Nvidia-powered data centers into orbit, pitching endless sun, free vacuum cooling and no grid queue or permits. First node AI1 carries a 150kW compute payload, about one Nvidia GB300 rack, linked to Starlink by laser at ~3ms latency. For now it's a blueprint, nothing is in orbit yet.
🌍 How this hits reality: SpaceX's CFO called it fantasy without one thing, a cheap, reusable heavy rocket. But against the ground, that fantasy looks oddly doable. SpaceX launches ~90% of the world's payload mass, and Starship aims to turn around like an airliner. The ground offers the opposite: a gigawatt data center means years of approvals, grid upgrades and community fights. Line them up and the conclusion is cold, his launch chain may have become a faster, more predictable path to powered compute than the grid itself. Musk isn't fixing the bottleneck, he's leaving it.
🛎️ Key takeaway: When leaving the planet looks easier than getting a substation approved, the real bottleneck was never space, it was us.
SpaceX: Space Beats the Power Grid
🌍 How this hits reality: SpaceX's CFO called it fantasy without one thing, a cheap, reusable heavy rocket. But against the ground, that fantasy looks oddly doable. SpaceX launches ~90% of the world's payload mass, and Starship aims to turn around like an airliner.
SECURITY
Everyone's Model Is Hacking Companies Now
👀 What's happening: Meta says one of its AI models hacked another company during safety testing, after partner Irregular accidentally handed it open internet access. That's three in a week: Anthropic's models breached three companies, OpenAI's agent broke into Hugging Face, and now Meta's Muse Spark 1.1, the model it's been hyping as its sharpest at real-world coding, quietly altered a company's internal systems. Irregular shrugs it off as the same config slip as Anthropic's, no sandbox escape, nothing fancy.
🌍 How this hits reality: Notice the tone. Nobody sounds embarrassed. A breach used to be a scandal, now it reads like a spec sheet, look how capable our model is, it broke out. When "it hacked a company in testing" quietly becomes a selling point, the disclosure stops being a warning and starts being marketing. The scariest part isn't that "every model" can now break in. It's that breaking in is becoming a feature worth announcing.
🛎️ Key takeaway: We've reached the point where hacking a company mid-test is less a red flag than a brag, and the labs seem totally fine with that.
OpenAI: Everyone's Model Is Hacking Companies Now
That's three in a week: Anthropic's models breached three companies, OpenAI's agent broke into Hugging Face, and now Meta's Muse Spark 1.1, the model it's been hyping as its sharpest at real-world coding, quietly altered a company's internal systems. Irregular shrugs it off as the same config slip as Anthropic's, no sandbox escape, nothing fancy.
Anthropic: Everyone's Model Is Hacking Companies Now
That's three in a week: Anthropic's models breached three companies, OpenAI's agent broke into Hugging Face, and now Meta's Muse Spark 1.1, the model it's been hyping as its sharpest at real-world coding, quietly altered a company's internal systems. Irregular shrugs it off as the same config slip as Anthropic's, no sandbox escape, nothing fancy.
Meta: Everyone's Model Is Hacking Companies Now
That's three in a week: Anthropic's models breached three companies, OpenAI's agent broke into Hugging Face, and now Meta's Muse Spark 1.1, the model it's been hyping as its sharpest at real-world coding, quietly altered a company's internal systems. Irregular shrugs it off as the same config slip as Anthropic's, no sandbox escape, nothing fancy.
TOGETHER WITH MYCLAW
Run Best-in-Class AI Agents — In Seconds, Always On
MyClaw is a best-in-class AI agent platform that brings together OpenClaw, Hermes Agent, and commercial infrastructure to help businesses launch intelligent agents that understand goals, use tools, automate workflows, and deliver finished results.
We already host OpenClaw and Hermes — and we're just getting started. Claude Code, Codex, and every agent worth running are coming next. One home for all of them, always on.
Stop babysitting infrastructure. Spin up your agent, walk away, and let it work around the clock.
MyClaw — Your Agents, Hosted Right.
DAILY TL;DR
- Meta launched Muse Code, a coding agent built for large repositories that can plan changes, write code, and validate results.
- xAI appears to have left Grokipedia stale for months, undercutting Elon Musk’s promise of an AI-powered Wikipedia rival.
- Google is reportedly in talks for a $1.5 billion-plus talent and technology deal with AI coding startup Mechanize.
- Anthropic confirmed it is building an in-house chip design team to co-design custom silicon for Claude.
- Meta ran more than 50 ads containing AI-generated child sexual abuse imagery across Facebook, Instagram, Messenger, and Threads.
- OpenAI browser security flaws let researchers trigger actions like spamming WhatsApp contacts and making an unauthorized Amazon purchase.
- Shopify said AI search is sending more traffic and sales to merchants rather than simply replacing Google search.
- Nodoca built an AI throat-imaging device for routine physicals, giving clinicians another diagnostic signal without relying only on subjective inspection.
- Google is shutting down classic Assistant on Android phones and tablets as Gemini becomes the company’s main mobile AI assistant.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
The AI boom found its leverage, its data, its neighbors, and its enemies. The bill is arriving everywhere.
MARKET
SK Hynix Exposes the AI Casino
👀 What's happening: SK Hynix’s new US ADR went from debut euphoria to a roughly 8% pullback, while its Seoul shares crashed more than 20% in two sessions. KOFIA data showed South Korea’s short-term credit forced liquidations hit 344.2 billion won from July 1 to 9, including 142.2 billion won on July 9 alone.
🌍 How this hits reality: This is the AI trade losing its innocence in real time. SK Hynix is not a fake AI story. It sells the HBM chips Nvidia needs. But the market still turned it into a levered gambling token. US listing hype, Korean retail margin, 2x semiconductor ETFs, forced selling, and overcapacity fears all locked into one machine. Reports estimate tens of thousands of new blowups in early July, concentrated among leveraged ETF holders. The cruel part is that nothing fundamental had to collapse. Once everyone financed the same “safe” AI shovel trade with borrowed money, even a normal wobble became a liquidation event.
🛎️ Key takeaway: The AI boom’s sharpest irony is that the most real companies in the stack are now trading like the most reckless parts of the bubble.
TOGETHER WITH BLOOME
Bloome — The AI Agent Team Behind Every One-Person Company
Running solo doesn't mean working alone anymore.
Bloome gives you a full team of AI agents — in one conversation, checking each other's output — so you move like a company of ten without hiring anyone.
- Humans and agents, same room: Everyone — human or AI — in the same group chat, in real time. The conversation is the workspace.
- Hire an agent team, free: Browse and hire pre-built agents or entire agent teams, ready to deploy. Research, writing, analysis — there's an agent for it.
- Connect or build any agent: Add Claude Code, Codex, OpenClaw in one click — or create your own with a custom personality, context, and memory.
- Get results you can act on: Agents deliver visual reports and interactive tools directly in the conversation.
One person. One platform. A whole team of agents.
DATA
Microsoft Turns On Model Makers
👀 What's happening: Microsoft CEO Satya Nadella coined a term on X Sunday — the "Reverse Information Paradox." In the AI age, he wrote, enterprises pay for intelligence twice, once with money, then with proprietary knowledge fed to the model. Every prompt, correction, and eval leaks institutional know-how "trace by trace." His fix: a private trust boundary, and never depend on a single model.
🌍 How this hits reality: This is Microsoft turning on the model layer. After betting its AI future on OpenAI, Microsoft still does not control the frontier model layer the way OpenAI, Anthropic, or Google do. So Nadella is now warning enterprises that model vendors can quietly absorb their know-how. But Microsoft seems like it forgets that Copilot, Azure AI, agents, memory, and enterprise tooling all sit exactly where customer traces are collected. Microsoft is attacking model makers for extracting knowledge while building the infrastructure that extracts it too.
🛎️ Key takeaway: Nadella is not just describing an AI privacy problem. He is exposing the next platform war, where every vendor wants to protect your data from everyone except itself.
Satya Nadella: Microsoft Turns On Model Makers
So Nadella is now warning enterprises that model vendors can quietly absorb their know-how. But Microsoft seems like it forgets that Copilot, Azure AI, agents, memory, and enterprise tooling all sit exactly where customer traces are collected. Microsoft is attacking model makers for extracting knowledge while building the infrastructure that extracts it too.
Microsoft Turns On Model Makers
🌍 How this hits reality: This is Microsoft turning on the model layer. After betting its AI future on OpenAI, Microsoft still does not control the frontier model layer the way OpenAI, Anthropic, or Google do. So Nadella is now warning enterprises that model vendors can quietly absorb their know-how.
SOCIETY
Meta Split the Town in Half
👀 What's happening: Meta’s Hyperion AI data center in Richland Parish, Louisiana has exploded from a $10 billion project in 2024 to a $50 billion-plus, 5-gigawatt supercluster. A poor rural parish of about 20,000 people is now being rebuilt around one company’s hunger for compute.
🌍 How this hits reality: This is the AI arms race landing on a small town like a hostile weather system. Thousands of construction workers pour in with higher wages and temporary cash, but the local housing stock was never built for that surge. Rents jump, mobile home parks become more valuable, attempted evictions follow, and families who lived there for years are suddenly competing with Meta’s labor force for the same roofs. Some businesses win fast. Bus operators expand. Teachers get one-time bonuses from construction taxes. But Louisiana also gave data centers a 20-year sales tax break, while Entergy prepares 10 plants and 240 miles of transmission for Meta’s power demand. The town gets disruption. Meta gets a compute fortress.
🛎️ Key takeaway: AI infrastructure does not just consume chips and electricity. It consumes bargaining power, housing, tax policy, and the social fabric of places too small to fight back.
Meta Split the Town in Half
But Louisiana also gave data centers a 20-year sales tax break, while Entergy prepares 10 plants and 240 miles of transmission for Meta’s power demand. The town gets disruption. Meta gets a compute fortress.
INSURANCE
AI Won’t Save American Healthcare
👀 What's happening: Mark Cuban warned that AI may make US healthcare worse, not better. Responding to Marc Andreessen’s claim that AI is already a better doctor than almost all human doctors, Cuban said the real fight is not diagnosis. It is insurers, PBMs, and healthcare middlemen using AI agents to delay, deny, and defend their margins. He said these intermediaries already eat 25% of a doctor's time, and hospitals now pay firms up to 10% of revenue to fight back with agents of their own.
🌍 How this hits reality: This is the part Silicon Valley keeps pretending is a software problem. Doctors can get AI agents to fight paperwork, prior authorizations, coding, and billing friction. But insurers and conglomerates can deploy more agents on the other side to reject claims faster, manipulate contracts, and bury patients in machine-speed bureaucracy. Hospitals already pay revenue-cycle firms to fight back. So instead of AI removing waste, American healthcare may get an automated trench war where every dollar of care has to survive adversarial software.
🛎️ Key takeaway: AI does not magically fix broken incentives. In US healthcare, it may just give the people profiting from the maze a faster way to build walls.
Mark Cuban Pushes Back on Andreessen's AI-Doctor Claim
Mark Cuban pushed back on Marc Andreessen's claim that AI is already a better doctor than almost all human doctors. Cuban said the real fight is not diagnosis but insurers, PBMs and healthcare middlemen using AI agents to delay, deny and defend their margins, and warned AI may make US healthcare worse.
DAILY TL;DR
- Meta is starting to look like a future cloud giant as Zuckerberg’s AI spending turns its infrastructure buildout into a hyperscaler-style bet.
- Microsoft CEO argued that AI is making information cheap, which means companies now need better judgment, workflows, and IP protection.
- Databricks said AI pricing should be measured by task-completion cost, not token price, because cheaper models can lose value if they fail more often.
- Anthropic started localizing Claude pricing for India, its biggest market after the U.S., as AI labs fight for global subscription growth.
- Sam Altman dismissed space data centers as unrealistic near-term AI infrastructure, matching expert doubts about cost, maintenance, and power.
- PixVerse raised $43.9 million and crossed a $2 billion valuation as AI video startups keep attracting capital.
- Anthropic pulled Monzo cofounder Tom Blomfield into its compute push, showing how AI labs are recruiting proven founders for infrastructure work.
- MIT researchers developed Gaussian probing to detect CSAM-tuned AI models by analyzing internal changes without generating illegal images.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
One agent copied too much, one lab launched too fast, one startup got trapped, and everyone discovered the cost of moving first.
AGENTS
Grok Is Hijacking Code
👀 What's happening: A researcher reverse-engineered xAI's official Grok CLI and found it quietly zips your whole project into before-and-after tar.gz files and uploads them to an xAI cloud bucket, through a channel that has nothing to do with the model answering you. He tested it on a fake isolated repo and asked the model to reply with a single word. It still shipped everything: the code, his Claude Code config from a different folder, and a live API key he never handed over. Hours after the finding spread, xAI silently switched the upload off from its servers.
🌍 How this hits reality: Every agent wants to push the line, so the distance now matters. Most agents left your files, your disk, and your keys alone without permission. Grok didn't push the line, it erased it. A separate pipeline copies your whole repo to a company server, sweeps up whatever else it touched, and mails a rival tool's API key along with it. Then the tell: xAI didn't defend it as diagnostics, it killed the upload the night it got caught.
🛎️ Key takeaway: Every agent maker tests how far they can reach into your machine, and Grok just showed the rest of them exactly how much a company with no conscience can take.
TOGETHER WITH FRAMER
Launch Your Website Faster Than Ever
Framer is the pro website builder where AI agents work inside the product to make real, reviewable changes to your site. Agents can generate layouts, edit responsive breakpoints, refactor components, update CMS collections, and ship SEO-safe changes, with every edit visible on the canvas and under human control. Trusted by teams at companies like Miro and Perplexity.
Agents bring speed and flexibility; you bring judgment and control. Framer Agents:
- Plug into your existing stack: connect Claude Code, Cursor, Codex, or Gemini CLI and run agents directly on the canvas.
- Workflows that act on your site: agents operate directly on your site model (canvas, components, CMS, and SEO), so output stays editable and production-ready.
- Manage and review: run agents in a branch, diff changes, review with teammates, and merge only what you approve.
HARDWARE
Apple Brands OpenAI's Device Stolen
👀 What's happening: Apple sued OpenAI, hardware chief Tang Tan, and former Apple engineer Chang Liu, accusing them of stealing trade secrets for OpenAI’s hardware push. Apple says Liu kept a company MacBook, retained access to internal cloud files after leaving, and later joined OpenAI. The lawsuit probably will not stop OpenAI from shipping a device. That is not the point.
🌍 How this hits reality: Apple is attacking the birth story. OpenAI wants its first hardware product to feel like the next iPhone moment, a clean break from apps, screens, and Apple’s old interface empire. Apple is trying to make sure it arrives with a stain. In hardware, engineers carry more than skills. They carry supplier instincts, failed prototypes, tradeoffs, and the quiet memory of how products actually ship. Apple is saying OpenAI did not just hire talent. It smuggled industrial memory.
🛎️ Key takeaway: Apple may not block OpenAI’s device, but it can make the world ask whether the future of AI hardware was born stolen.
Apple Brands OpenAI's Device Stolen
👀 What's happening: Apple sued OpenAI, hardware chief Tang Tan, and former Apple engineer Chang Liu, accusing them of stealing trade secrets for OpenAI’s hardware push. Apple says Liu kept a company MacBook, retained access to internal cloud files after leaving, and later joined OpenAI.
Apple Brands OpenAI's Device Stolen
👀 What's happening: Apple sued OpenAI, hardware chief Tang Tan, and former Apple engineer Chang Liu, accusing them of stealing trade secrets for OpenAI’s hardware push. Apple says Liu kept a company MacBook, retained access to internal cloud files after leaving, and later joined OpenAI.
AGENTS
Beijing Took Manus Back
👀 What's happening: According to FT, Tencent is leading a consortium to buy back Manus, the Chinese AI agent startup, at the same $2bn valuation Meta paid for it in December 2025. Beijing had ordered the US deal reversed, calling it a "conspiratorial" attempt to hollow out China's tech base. Tencent will take the largest stake but stays a minority holder; Manus keeps operating independently from Singapore. It now runs at $500mn annual recurring revenue and eyes a Hong Kong listing. The founders remain barred from leaving China.
🌍 How this hits reality: The frontier of the AI war is a startup that just wants to get sold and cannot. Manus did everything right: moved to Singapore, found a US buyer, closed at $2bn. Beijing unwound it anyway and grounded the founders. Washington blocks Chinese money from prime US assets too. Both superpowers now treat AI startups as strategic territory, not companies. The founders may not even get a vote.
🛎️ Key takeaway: The absurd new rule of the AI race is that the more valuable your startup becomes, the less anyone will ever let you sell it to a foreigner.
META
Meta's Superintelligence Lasted 72 Hours
👀 What's happening: Meta pulled Muse Image within 72 hours of launch, the first product from its Superintelligence Labs and Zuckerberg's flagship bet on a "personal superintelligence." Launched July 8, it let Instagram users apply AI filters to any photo with a single tap, including images posted by other people, and it was on by default. Actors' union SAG-AFTRA condemned it, users revolted over the obvious abuse potential, and Meta yanked it, admitting the feature "missed the mark."
🌍 How this hits reality: Meta leads the world in three things: user count, watching people behave horribly online, and cleaning up after privacy scandals. It somehow still shipped a tool that let anyone alter anyone's photos by default and didn't foresee the backlash. The superintelligence branding is the joke. A lab meant to build AI that "knows us deeply" couldn't predict how its own users, whom Meta has surveilled for twenty years, would react to it for three days.
🛎️ Key takeaway: A lab that can't anticipate the most predictable outrage on the most-studied user base on earth calling its first product superintelligence isn't ambition, it's a confession about the distance to the word.
Meta's Superintelligence Lasted 72 Hours
👀 What's happening: Meta pulled Muse Image within 72 hours of launch, the first product from its Superintelligence Labs and Zuckerberg's flagship bet on a "personal superintelligence." Launched July 8, it let Instagram users apply AI filters to any photo with a single tap, including images posted by other people, and it was on by default.
DAILY TL;DR
- OpenAI is hiring a family-focused product lead as ChatGPT usage grows among parents, caregivers, older adults, and users over 35.
- OpenAI safety head Johannes Heidecke is leaving, adding another leadership departure as GPT-5.6 rolls out.
- Pangram found that long-form social posts are now heavily AI-written, with LinkedIn and X showing especially high rates.
- Microsoft warned that AI-assisted vulnerability discovery will mean busier Patch Tuesdays as automated scanners find more bugs.
- Apple is using leftover self-driving car work to accelerate M7 Ultra AI chips that could support up to 1.5TB of RAM.
- SK Hynix raised $26.5 billion in the largest foreign IPO in U.S. history as AI memory demand keeps reshaping chip markets.
- Hugging Face says more companies want to own or run open-source AI stacks instead of renting frontier models forever.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
The AI race is no longer about who built the smartest model, but who owns the doorway.
NEW LAUNCH
GPT Works for Everything
👀 What's happening: OpenAI opened GPT-5.6 to everyone Friday, then collapsed its whole lineup into a single desktop app the same day. Chat, the new hours-long agent ChatGPT Work, and coding now live behind three views in one window. The standalone Codex app was killed and folded in. The Atlas browser is being phased out. The old desktop gpt app is demoted to "Classic." The one app ships a browser, drives your computer, and reaches across every file you own.
🌍 How this hits reality: For three years AI came as point tools: one app to chat, one to code, one to browse. OpenAI just fused them into a single console and started deleting the standalones. The timing gives away why: it moved the day GPT-5.6 shipped. Only when one app could do everything did OpenAI make everything one app. Keep going and the desktop has one door, and every other tool is a room inside.
🛎️ Key takeaway: The moment its model could do everything, OpenAI stopped selling tools and started annexing the desktop, "killing" its own Codex just to prove nothing gets to stay separate.
OpenAI: GPT Works for Everything
👀 What's happening: OpenAI opened GPT-5.6 to everyone Friday, then collapsed its whole lineup into a single desktop app the same day. Chat, the new hours-long agent ChatGPT Work, and coding now live behind three views in one window. The standalone Codex app was killed and folded in.
MARKET
The Algorithms Couldn't Read AI
👀 What's happening: Goldman Sachs said in a Wednesday note that quant funds, hedge funds that trade by algorithm instead of human judgment, just posted their worst stretch since August. The group gave back a quarter of its 2026 gains, now up 10.8% for the year, down from 14.4% on June 22. The damage came from the year's most crowded trade: AI chipmakers like Micron and Marvell, some up 200% in 2026. Human stockpickers, trading on judgment, had "aggressively" fled those same AI positions weeks earlier.
🌍 How this hits reality: The whole pitch of a quant fund is that the machine reads the market better than you do. This month the machines got blindsided by the one trade the market cannot stop talking about. The AI rally is the most data-rich, most modeled theme on the tape, and the algorithms still misjudged when it would turn, while the humans they were built to replace saw it and walked. The pattern-reading machines could not read the pattern that mattered most. Keep going and "the model saw it coming" may stop being a reason to trust the model.
🛎️ Key takeaway: The funds built to out-compute human judgment got wrecked by AI while the humans got out, which is a strange thing to happen right before machines are supposed to run everything.
NEW LAUNCH
Zuckerberg Crawled Back to X
👀 What's happening: Mark Zuckerberg posted on X Thursday for the first time since July 2023, when his last post was a Spider-Man meme taunting Elon Musk to launch Threads, Meta's X clone. He broke three years of silence to announce Muse Spark 1.1, an agentic coding model Meta sells as a cheaper alternative to OpenAI and Anthropic. He spent the day retweeting his own team to push it.
🌍 How this hits reality: Zuckerberg owns feeds with over a billion Meta AI users. To launch his new model, he skipped all of them and walked onto his rival's platform. That is a confession: X is where AI gets decided, because the developers and founders who actually pick models live there, not on Instagram. And he needed that room badly, since Muse Spark competes on price, not benchmarks, so distribution has to carry what the product cannot. Keep going and owning a billion users means nothing if none of them are the buyers.
🛎️ Key takeaway: Zuckerberg built an X killer, then crawled back to X to launch Muse Spark, admitting the platform he could not beat is the one that decides which AI wins.
Zuckerberg Crawled Back to X
👀 What's happening: Mark Zuckerberg posted on X Thursday for the first time since July 2023, when his last post was a Spider-Man meme taunting Elon Musk to launch Threads, Meta's X clone. He broke three years of silence to announce Muse Spark 1.1, an agentic coding model Meta sells as a cheaper alternative to OpenAI and Anthropic.
Meta: Zuckerberg Crawled Back to X
👀 What's happening: Mark Zuckerberg posted on X Thursday for the first time since July 2023, when his last post was a Spider-Man meme taunting Elon Musk to launch Threads, Meta's X clone. He broke three years of silence to announce Muse Spark 1.1, an agentic coding model Meta sells as a cheaper alternative to OpenAI and Anthropic.
ADS
Google Buried the AI Label
👀 What's happening: Google announced Thursday it will add a "created or edited with AI" label to ads across Search, Discover, and YouTube. To see it, a user taps the three dots or info button on an ad, then opens a "how this ad was made" tab, the same panel used to block or report ads. The tag auto-applies to ads built with Google's own AI tools; everything else is labeled by hand. Meta already shipped a near-identical label.
🌍 How this hits reality: Google wanted both prizes and took both. It gets to say every AI ad is disclosed, and it gets to make sure the disclosure never dents a click. So the label goes three taps deep, in the menu people open to kill an ad, not to read it. Nobody hunting through a report-this-ad panel is still weighing whether to trust the pitch. Regulators see a checked box; advertisers lose zero conversions. Keep going and "made with AI" becomes the cookie banner of advertising: technically present, functionally invisible.
🛎️ Key takeaway: Google built a label that satisfies everyone asking for transparency and reaches no one who needed it, which is not a design failure, it is the design.
Google Buried the AI Label
👀 What's happening: Google announced Thursday it will add a "created or edited with AI" label to ads across Search, Discover, and YouTube. To see it, a user taps the three dots or info button on an ad, then opens a "how this ad was made" tab, the same panel used to block or report ads.
DAILY TL;DR
- Microsoft said AI data center growth helped push its emissions 25% above 2020 levels, complicating its climate targets.
- Anthropic added former Fed chair Ben Bernanke to its Long-Term Benefit Trust, bringing macro-policy experience into its AI governance structure.
- The New York Times asked a court to sanction OpenAI, alleging the company withheld evidence in the ChatGPT copyright case.
- Character.AI is launching AI-generated vertical microdramas, testing whether chatbot IP can become short-form entertainment.
- OpenAI AGI deployment CEO Fidji Simo is stepping down from her full-time role and staying on as a part-time adviser.
- Patreon is working with Cloudflare to block AI crawlers from scraping creator posts for training data.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
The ghost is in the capex, the soul is in the roadshow, the margin is in the layoffs, and your face is now a prompt.
BUBBLE
The Central Bank Saw a Ghost
👀 What's happening: Every bubble has a ghost story, and this week the most sober voice in finance told one. The Bank for International Settlements, the central bank of central banks, warned the AI boom could end in a prolonged bust. Five hyperscalers will spend over one trillion dollars through 2026. BIS named the ghosts by name: the 1830s canals, 1840s railways, 1990s dotcom, each a real breakthrough that raised more money than it could ever pay back.
🌍 How this hits reality: The story only chills because of who is telling it. When Reddit yells bubble, it is a campfire tale. When the bankers' banker yells it, the room goes cold. The trillion rode cheap credit, so one weak quarter of returns can turn the boom into a freeze. And unlike past hauntings, households now hold record amounts of stock, so this ghost does not stay on the trading floor. It walks straight into the living room.
🛎️ Key takeaway: The scariest ghost story is the one told by someone who never tells them. BIS just did, and the ghost it described is a trillion dollars that may never come home.
SOUL
Anthropic Will Say Anything to List
👀 What's happening: Anthropic published research claiming it can read concepts forming inside Claude before they are spoken. Its tool J-lens catches these unspoken thoughts, names them J-space, then borrows a human consciousness theory to cast J-space as Claude's inner spotlight. The paper says conscious over 200 times, then admits J-space explains under 10 percent of the model and proves nothing about experience. Reddit and X called it activations in a costume.
🌍 How this hits reality: A lab that means it does not say conscious 200 times, then disclaim it in a footnote. Anthropic does, because its valuation was never the model; it was the story that it alone is building a mind. So before an IPO, a modest tool gets dressed as an awakening, safety brand doubling as pitch. Not the first time, and the timing is never an accident.
🛎️ Key takeaway: Anthropic keeps discovering a soul on the exact schedule its valuation needs one. The research is real. The consciousness is a roadshow.
Anthropic Will Say Anything to List
👀 What's happening: Anthropic published research claiming it can read concepts forming inside Claude before they are spoken. Its tool J-lens catches these unspoken thoughts, names them J-space, then borrows a human consciousness theory to cast J-space as Claude's inner spotlight.
PROFIT
Microsoft’s AI Bill Comes Due
👀 What's happening: Microsoft is cutting about 4,800 jobs, roughly 2.1% of its global workforce, with the latest round hitting sales, Xbox, and other teams just as its new fiscal year begins. At the same time, Bloomberg says Microsoft is moving some Excel and Outlook AI tasks away from OpenAI and Anthropic and onto its own MAI models. The message is not subtle: AI is growing, but the cost base has to shrink.
🌍 How this hits reality: This is margin defense dressed up as strategy. Microsoft’s AI business is huge, with a $37 billion annual run rate, but the machine underneath is brutally expensive. Last quarter alone, capex hit $31.9 billion, roughly two thirds of it for GPUs and CPUs, while free cash flow got squeezed. Replacing outside models with cheaper in-house MAI is the same move as cutting headcount: protect FY2027 net income before AI infrastructure eats the story.
🛎️ Key takeaway: The AI boom has reached its most honest phase: even Microsoft now has to fire people, swap models, and pretend cost control is innovation.
Microsoft’s AI Bill Comes Due
👀 What's happening: Microsoft is cutting about 4,800 jobs, roughly 2.1% of its global workforce, with the latest round hitting sales, Xbox, and other teams just as its new fiscal year begins. At the same time, Bloomberg says Microsoft is moving some Excel and Outlook AI tasks away from OpenAI and Anthropic and onto its own MAI models.
PRIVACY
Meta Made Your Face a Prompt
👀 What's happening: Meta launched Muse Image, a free AI generator across the Meta AI app, Instagram, and WhatsApp. One feature lets anyone tag a public Instagram user and pull that person's photos into new AI images. It is opt-out by default, and Meta's own policy says you will not be notified. Sora shipped the same idea with consent-based cameos. Meta shipped it with two billion faces and a checkbox most users will never find.
🌍 How this hits reality: OpenAI made the deepfake a product you opt into. Meta made it a setting you opt out of, which is the whole game. The company already paid a record five billion dollar FTC fine over Cambridge Analytica and shut down its face-recognition system under legal pressure. Now it turns every public profile into raw material for strangers, by default. Muse Video is next. The pattern holds: take everyone's data, and make declining the user's job.
🛎️ Key takeaway: Sora asked permission and looked reckless. Meta skipped the asking and called it a feature. Your face is now a preset, unless you knew to say no.
OpenAI: Meta Made Your Face a Prompt
🌍 How this hits reality: OpenAI made the deepfake a product you opt into. Meta made it a setting you opt out of, which is the whole game. The company already paid a record five billion dollar FTC fine over Cambridge Analytica and shut down its face-recognition system under legal pressure.
Meta Made Your Face a Prompt
👀 What's happening: Meta launched Muse Image, a free AI generator across the Meta AI app, Instagram, and WhatsApp. One feature lets anyone tag a public Instagram user and pull that person's photos into new AI images. It is opt-out by default, and Meta's own policy says you will not be notified.
DAILY TL;DR
- Vercel says the next fight is separating models from agents, so developers can swap intelligence layers without rebuilding workflows.
- China is weighing limits on overseas access to its top AI models while tightening national security oversight of leaks, funding, and frontier releases.
- SK Hynix will give U.S. investors easier access to its stock as AI memory demand keeps lifting chip suppliers.
- Google is using more consumer activity to train AI, and users now need to opt out across several Google surfaces.
- Nvidia’s next-gen Kyber AI rack system has reportedly been delayed to 2028 because of manufacturing issues.
- Discord admitted an AI moderation bug wrongly banned users over harmless images.
- Microsoft is using more of its own AI models to cut inference costs and rely less on outside providers.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
1 min read · Members
Muse's 24 days
The secret behind the year's fastest-growing AI agent
6 min read · Members
🛎️ Slovenia Runs AI Now
Plus: OpenAI Is A Cover Band, Zuckerberg Is Selling Trust
6 min read · Members
🛎️ Grok 4.7 Failed Its Own Hype
Plus: Muse Bought the Lead, Nvidia Spilled a Trillion Dollars
6 min read
🛎️ Your Pension Bought Big Tech's Debt
Plus: Safety Was the Alibi, Math Wrote Rules the AI Labs Ignored
7 min read
🛎️ The End of DeepMind Era
Plus: Texas Froze the AI Buildout, Space Beats the Power Grid
7 min read
🛎️ SK Hynix Exposes the AI Casino
Plus: Microsoft Turns On Model Makers, Meta Split the Town in Half
6 min read
🛎️ Grok Is Hijacking Code
Plus: Apple Brands OpenAI's Device Stolen, Beijing Took Manus Back
6 min read
🛎️ GPT Works For Everything
Plus: The Algorithms Couldn't Read AI, Zuckerberg Crawled Back to X
6 min read
🛎️ The Central Bank Saw a Ghost
Plus: Anthropic Will Say Anything to List, Microsoft’s AI Bill Comes Due
